Learning a deterministic finite automaton with a recurrent neural network

Laura Firoiu, Tim Oates, Paul R Cohen

Research output: Chapter in Book/Report/Conference proceedingConference contribution

4 Citations (Scopus)

Abstract

We consider the problem of learning a finite automaton with recurrent neural networks from positive evidence. We train an Elman recurrent neural network with a set of sentences in a language and extract a finite automaton by clustering the states of the trained network. We observe that the generalizations beyond the training set, in the language recognized by the extracted automaton, are due to the training regime: the network performs a “loose” minimization of the prefix DFA of the training set, the automaton that has a state for each prefix of the sentences in the set.

Original languageEnglish (US)
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
PublisherSpringer Verlag
Pages90-101
Number of pages12
Volume1433
ISBN (Print)3540647767, 9783540647768
StatePublished - 1998
Externally publishedYes
Event4th International Colloquium on Grammatical Inference, ICGI 1998 - Ames, United States
Duration: Jul 12 1998Jul 14 1998

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume1433
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other4th International Colloquium on Grammatical Inference, ICGI 1998
CountryUnited States
CityAmes
Period7/12/987/14/98

Fingerprint

Deterministic Finite Automata
Recurrent neural networks
Recurrent Neural Networks
Finite automata
Finite Automata
Prefix
Automata
Elman Neural Network
Clustering
Learning
Training
Language

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Firoiu, L., Oates, T., & Cohen, P. R. (1998). Learning a deterministic finite automaton with a recurrent neural network. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 1433, pp. 90-101). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1433). Springer Verlag.

Learning a deterministic finite automaton with a recurrent neural network. / Firoiu, Laura; Oates, Tim; Cohen, Paul R.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 1433 Springer Verlag, 1998. p. 90-101 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1433).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Firoiu, L, Oates, T & Cohen, PR 1998, Learning a deterministic finite automaton with a recurrent neural network. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 1433, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 1433, Springer Verlag, pp. 90-101, 4th International Colloquium on Grammatical Inference, ICGI 1998, Ames, United States, 7/12/98.
Firoiu L, Oates T, Cohen PR. Learning a deterministic finite automaton with a recurrent neural network. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 1433. Springer Verlag. 1998. p. 90-101. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
Firoiu, Laura ; Oates, Tim ; Cohen, Paul R. / Learning a deterministic finite automaton with a recurrent neural network. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 1433 Springer Verlag, 1998. pp. 90-101 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{a249e23c4f88442093c4a3ef20705249,
title = "Learning a deterministic finite automaton with a recurrent neural network",
abstract = "We consider the problem of learning a finite automaton with recurrent neural networks from positive evidence. We train an Elman recurrent neural network with a set of sentences in a language and extract a finite automaton by clustering the states of the trained network. We observe that the generalizations beyond the training set, in the language recognized by the extracted automaton, are due to the training regime: the network performs a “loose” minimization of the prefix DFA of the training set, the automaton that has a state for each prefix of the sentences in the set.",
author = "Laura Firoiu and Tim Oates and Cohen, {Paul R}",
year = "1998",
language = "English (US)",
isbn = "3540647767",
volume = "1433",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "90--101",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - Learning a deterministic finite automaton with a recurrent neural network

AU - Firoiu, Laura

AU - Oates, Tim

AU - Cohen, Paul R

PY - 1998

Y1 - 1998

N2 - We consider the problem of learning a finite automaton with recurrent neural networks from positive evidence. We train an Elman recurrent neural network with a set of sentences in a language and extract a finite automaton by clustering the states of the trained network. We observe that the generalizations beyond the training set, in the language recognized by the extracted automaton, are due to the training regime: the network performs a “loose” minimization of the prefix DFA of the training set, the automaton that has a state for each prefix of the sentences in the set.

AB - We consider the problem of learning a finite automaton with recurrent neural networks from positive evidence. We train an Elman recurrent neural network with a set of sentences in a language and extract a finite automaton by clustering the states of the trained network. We observe that the generalizations beyond the training set, in the language recognized by the extracted automaton, are due to the training regime: the network performs a “loose” minimization of the prefix DFA of the training set, the automaton that has a state for each prefix of the sentences in the set.

UR - http://www.scopus.com/inward/record.url?scp=79960187948&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79960187948&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:79960187948

SN - 3540647767

SN - 9783540647768

VL - 1433

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 90

EP - 101

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

PB - Springer Verlag

ER -